
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px
import re
import os
%matplotlib inline

from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import MinMaxScaler
from sklearn.decomposition import PCA
from sklearn.cluster import KMeans
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import MinMaxScaler

import warnings
warnings.filterwarnings('ignore', category=FutureWarning)

pd.options.mode.chained_assignment = None # Disable those setting with copy warnings


def set_style(df, caption=""):
    """Helper function to set dataframe presentation style.
    """
    return df.style.background_gradient(cmap='Purples').set_caption(caption).set_table_styles([{
    'selector': 'caption',
    'props': [
        ('color', 'Purple'),
        ('font-size', '18px'),
        ('font-weight','bold')
    ]}])

def bar_annot(ax, annot_format='{:.2f}'):
    """Function to annotate bar plots. ax is the barplot axis object. annot_format is the annotation numeric display format.
    """
    for p in ax.patches:
        annot = annot_format.format(p.get_height())
        ax.annotate(annot, (p.get_x(), p.get_height()), weight='bold')

def convert_time(x)->int:
    """run_time column has xxhyym or xxh or yym format. This method converts that into integer minutes. Default return is 0.
    """
    x = x.replace(' ','') # Remove whitespace for processing
    
    if re.match("^([0-9]+)h([0-9]+)m$", x):
        # xxhyym format
        l = x.split('h')
        h = int(l[0])
        ll = l[1].split('m')
        m = int(ll[0])
        return h*60 + m
    elif re.match("^([0-9]+)h$", x):
        # xxh format
        l = x.split('h')
        h = int(l[0])
        return h * 60
    elif re.match("^([0-9]+)m$", x):
        # yym format
        l = x.split('m')
        m = int(l[0])
        return m
    else:
        # Anything else
        return 0


dataset = pd.read_csv('dataset_imdb_original.csv')
dataset.head()


print("Data Size:",dataset.shape)
print("-"* 30)
print("About Dataset:")
dataset.info()
print("-"* 30)
print("Data Columns:",list(dataset.columns))
print("-"* 30)
print("Number of Examples 'N'=",dataset.shape[0])
print("Number of Dimensions 'D'=",dataset.shape[1]-1)
print("-"* 30)
print("Data Check For Any Duplicates:",dataset.duplicated().any())

Data Size: (250, 13)
------------------------------
About Dataset:
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 250 entries, 0 to 249
Data columns (total 13 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   rank         250 non-null    int64  
 1   name         250 non-null    object 
 2   year         250 non-null    int64  
 3   rating       250 non-null    float64
 4   genre        250 non-null    object 
 5   certificate  250 non-null    object 
 6   run_time     250 non-null    object 
 7   tagline      250 non-null    object 
 8   budget       250 non-null    object 
 9   box_office   250 non-null    object 
 10  casts        250 non-null    object 
 11  directors    250 non-null    object 
 12  writers      250 non-null    object 
dtypes: float64(1), int64(2), object(10)
memory usage: 25.5+ KB
------------------------------
Data Columns: ['rank', 'name', 'year', 'rating', 'genre', 'certificate', 'run_time', 'tagline', 'budget', 'box_office', 'casts', 'directors', 'writers']
------------------------------
Number of Examples 'N'= 250
Number of Dimensions 'D'= 12
------------------------------
Data Check For Any Duplicates: False


dataset.describe()


notavailabe_values = (dataset == 'Not Available').sum()
print(notavailabe_values)

rank            0
name            0
year            0
rating          0
genre           0
certificate     1
run_time        1
tagline         0
budget         39
box_office     30
casts           0
directors       0
writers         0
dtype: int64


dataset['budget'] = dataset['budget'].str.replace('$','')
dataset['budget'] = dataset['budget'].str.replace('Not Available','0')
dataset['budget'] = dataset['budget'].apply(lambda x: float(''.join(filter(str.isdigit, str(x)))))
# Convert the "budget" column to a numeric data type, and replace "Not Available" values with NaN
dataset['budget'] = pd.to_numeric(dataset['budget'], errors='coerce', downcast='float')
dataset['budget'] = dataset['budget'].fillna(dataset['budget'].median())


dataset['box_office'] = dataset['box_office'].str.replace('$','')
dataset['box_office'] = dataset['box_office'].str.replace('Not Available','0')
dataset['box_office'] = dataset['box_office'].apply(lambda x: float(''.join(filter(str.isdigit, str(x)))))
dataset['box_office'] = pd.to_numeric(dataset['box_office'], errors='coerce', downcast='float')
dataset['box_office'] = dataset['box_office'].fillna(dataset['box_office'].median())


dataset['return_rate'] = dataset['box_office'] / dataset['budget'] * 100


def runtime_to_minutes(runtime_str):
    if runtime_str == "Not Available":
        return np.nan
    elif "h " in runtime_str:
        hours, minutes = runtime_str.split("h ")
        return int(hours) * 60 + int(minutes[:-1])
    else:
        return int(runtime_str[:-1])

# Apply the function to the run_time column
dataset['runtime_minutes'] = dataset['run_time'].apply(runtime_to_minutes)


dataset.describe()


genre_counts = dataset['genre'].value_counts()

display(genre_counts.head(5))

real_gen = ""
for i in dataset['genre']:
    real_gen = real_gen+","+i
all_gen = real_gen.split(',')[1:]
unique_genres = list(set(all_gen))

data = []
for i in unique_genres:
    c = 0
    for j in dataset['genre']:
        if i in j:
            c+=1
    data.append([i, c])
    
gen_df = pd.DataFrame(data, columns=['Genre', 'Freq'])
gen_df.sort_values(by = 'Freq', ascending = False, inplace = True)
gen_df = gen_df.head(5)
plt.figure(figsize=(12,6),dpi = 150)
sns.barplot(data = gen_df, x = 'Genre', y = 'Freq', palette='Paired')
plt.title('Top 5 genêros mais populares')
plt.show()

Drama                         19
Crime,Drama                   14
Animation,Adventure,Comedy    10
Biography,Drama,History        9
Crime,Drama,Mystery            9
Name: genre, dtype: int64


# Compute the mean budget excluding NaN Values
average_budget = dataset['budget'].mean(skipna=True)

print(f"O orçamento médio para um filme no conjunto de dados é: {average_budget:.2f}")

O orçamento médio para um filme no conjunto de dados é: 44257860.00


# Calculate the average box office revenue for movies in the dataset
avg_box_office = dataset['box_office'].mean()

print(f"A receita média de bilheteria para um filme no conjunto de dados é: {avg_box_office:.2f}")

A receita média de bilheteria para um filme no conjunto de dados é: 206773984.00


# Calculate the average runtime in miunutes
avg_runtime = dataset['runtime_minutes'].mean()

# Convert the average runtime back to hours and minutes format
hours = int(avg_runtime/60)
minutes = int(avg_runtime % 60)
print(f"O tempo de execução médio de um filme no conjunto de dados é: {hours}h {minutes}m.")

O tempo de execução médio de um filme no conjunto de dados é: 2h 6m.


# Filter out the 'Not Available' box office values and sort by descending order
top_grossing = dataset[dataset['box_office'] != 'Not Available'].sort_values(by='box_office', ascending=False).head(10)

# Create a list of 10 different colors
colors = ['#F44336', '#E91E63', '#9C27B0', '#673AB7', '#3F51B5', '#2196F3', '#00BCD4', '#4CAF50', '#8BC34A', '#FFC107']

# Plot the bar chart
plt.figure(figsize=(16,6))
plt.bar(x=top_grossing['name'], height=top_grossing['box_office'], color=colors)
plt.xticks(rotation=90)
plt.xlabel('Nome do Filme')
plt.ylabel('Bilheteria ($)')
plt.title('Top 10 Maior Bilheteria')
plt.show()


plt.figure(figsize=(16,6))
dataset.groupby('directors').size().sort_values(ascending=False).head(10).plot(kind = 'bar', color = sns.color_palette('Set3'))
plt.xlabel("Nome dos Diretores")
plt.ylabel("Número de filmes")
plt.title("Top 10 diretores que dirigiram mais filmes")
plt.show()


director_ratings = dataset.groupby("directors")["rating"].mean()
director_ratings = director_ratings.sort_values(ascending=False)

top_directors = director_ratings.head(10)

plt.figure(figsize=(16,6))
director_ratings.head(10).plot(kind = 'bar', color = sns.color_palette('Set3'))
plt.xlabel("Classificação média")
plt.ylabel("Diretor")
plt.title("Os 10 melhores diretores por classificação média no conjunto de dados")
plt.show()


na_counts = dataset.apply(lambda x: x.value_counts().get('Not Available', 0))
print('Not Available Values:')
print(na_counts)

Not Available Values:
rank               0
name               0
year               0
rating             0
genre              0
certificate        1
run_time           1
tagline            0
budget             0
box_office         0
casts              0
directors          0
writers            0
return_rate        0
runtime_minutes    0
dtype: int64


numerical_data = dataset.select_dtypes(include='number')

for column in numerical_data:
    print("'" + column + "'", "Range Of Values:",numerical_data[column].min(),"-",numerical_data[column].max())
    print("-"* 30)

'rank' Range Of Values: 1 - 250
------------------------------
'year' Range Of Values: 1921 - 2022
------------------------------
'rating' Range Of Values: 8.0 - 9.3
------------------------------
'budget' Range Of Values: 0.0 - 2400000000.0
------------------------------
'box_office' Range Of Values: 0.0 - 2799439104.0
------------------------------
'return_rate' Range Of Values: 0.002093750052154064 - inf
------------------------------
'runtime_minutes' Range Of Values: 2.0 - 238.0
------------------------------


corr = round(dataset.corr(), 2)
sns.heatmap(corr, annot= True, cmap='viridis')

<AxesSubplot:>


z_scores = numerical_data[numerical_data.columns[1:]].apply(lambda x: (x - x.mean()) / x.std())
outliers = (z_scores > 3) | (z_scores < -3)
print(outliers.sum())

year               0
rating             7
budget             2
box_office         5
return_rate        0
runtime_minutes    6
dtype: int64


top_movies_by_genre = dataset.groupby("genre").apply(lambda x: x.loc[x["rating"].idxmax()])
top_movies_by_genre[["rating", "name"]].head()


plt.hist(dataset["rating"], bins=20, edgecolor='black', color='green')
plt.xlabel("Avaliação")
plt.ylabel("Número de filmes")
plt.title("Distribuição de classificações no conjunto de dados")
plt.show()


casts_ratings = dataset.groupby("casts")["rating"].mean()
casts_ratings = casts_ratings.sort_values(ascending=False)
top_casts = casts_ratings.head(10)
shortened_labels = [name[:10] + '...' if len(name) > 10 else name for name in top_casts.index]
plt.figure(figsize=(16,6))
plt.barh(top_casts.index, top_casts.values, edgecolor='black', color='green')
plt.yticks(top_casts.index, shortened_labels)
plt.xlabel("Classificação média")
plt.ylabel("Elencos")
plt.title("Os 10 melhores elencos por classificação média")
plt.show()


dict_genres = {}
for genres in dataset.genre:
    genre_list = genres.split(',')
    for g in genre_list:
        genre = g.strip()
        if genre not in dict_genres:
            dict_genres[genre] = 1
        else:
            dict_genres[genre] += 1

print(dict_genres)

{'Drama': 177, 'Crime': 51, 'Action': 50, 'Biography': 29, 'History': 10, 'Adventure': 60, 'Western': 7, 'Romance': 23, 'Sci-Fi': 20, 'Fantasy': 14, 'Mystery': 31, 'Family': 13, 'Thriller': 30, 'War': 23, 'Comedy': 45, 'Animation': 23, 'Music': 4, 'Horror': 5, 'Film-Noir': 4, 'Musical': 1, 'Sport': 5}


year_counts = dataset['year'].value_counts()

year_counts = dataset.groupby('year')['name'].count()

sns.set(style="darkgrid")
plt.plot(year_counts.index, year_counts.values)
plt.xlabel('Ano')
plt.ylabel('Número de filmes lançados')
plt.title('Número de filmes lançados a cada ano')
plt.show()


year_ratings = dataset.groupby("year")["rating"].mean()
plt.plot(year_ratings.index, year_ratings.values, color='purple')
plt.xlabel("Ano")
plt.ylabel("Classificação média")
plt.title("Análise anual dos filmes com melhor classificação no conjunto de dados")
plt.show()


scaler = MinMaxScaler()
scaler.fit(dataset[['budget', 'box_office']])
scaled_data = scaler.transform(dataset[['budget', 'box_office']])

genres = dataset['genre'].tolist()
certificates = dataset['certificate'].tolist()
taglines = dataset['tagline'].tolist()

tfidf_genres = TfidfVectorizer()
tfidf_certificate = TfidfVectorizer()
tfidf_tagline = TfidfVectorizer()

tfidf_genres.fit(genres)
tfidf_certificate.fit(certificates)
tfidf_tagline.fit(taglines)

feature_names_genres = tfidf_genres.get_feature_names_out()
feature_names_certificates = tfidf_certificate.get_feature_names_out()
feature_names_taglines = tfidf_tagline.get_feature_names_out()

def Text_Score(text, tfidf, feature_names):
    tfidf_matrix= tfidf.transform([text]).todense()
    feature_index = tfidf_matrix[0,:].nonzero()[1]
    tfidf_scores = zip([feature_names[i] for i in feature_index], [tfidf_matrix[0, x] for x in feature_index])
    text_dict = dict(tfidf_scores)
    text_scores = text_dict.values()
    return sum(text_scores)
dataset['genre_score'] = dataset['genre'].apply(lambda x: Text_Score(x, tfidf_genres, feature_names_genres))
dataset['certificate_score'] = dataset['certificate'].apply(lambda x: Text_Score(x, tfidf_certificate, feature_names_certificates))
dataset['tagline_score'] = dataset['tagline'].apply(lambda x: Text_Score(x, tfidf_tagline, feature_names_taglines))


plt.figure(figsize=(14,7))
dataset.groupby('certificate').size().sort_values(ascending=False).head(5).plot(kind = 'bar', color = sns.color_palette('Set2'))
plt.xlabel('Certificado')
plt.ylabel('Frequência')
plt.title("5 principais certificados")
plt.show()


dataset['year'].sort_values()

_21to40 = 0
_41to60 = 0
_61to80 = 0
_81to00 = 0
_01to22 = 0

for i in dataset['year']:
    if i>1920 and i<=1940:
        _21to40+=1
    elif i>1940 and i<=1960:
        _41to60+=1
    elif i>1960 and i<=1980:
        _61to80+=1
    elif i>1980 and i<=2000:
        _81to00+=1
    else: 
        _01to22+=1
        
data = [_21to40, _41to60, _61to80, _81to00, _01to22]
label = ['1921 - 1940', '1941- 1960', '1961 - 1980', '1981 - 2000', '2001 - 2022']
plt.figure(figsize=(12,6))
plt.pie(data,labels=label,colors=sns.color_palette('Paired'),autopct='%.2f%%')
plt.title("Porcentagem de filmes com base no ano")
plt.show()


fig = px.box(dataset, x='genre', y='box_office', color='genre', title='Bilheteria por Gênero')
fig.show()


fig = px.violin(dataset, x='certificate', y='box_office', color='genre', title='Bilheteira por Certificado')
fig.show()


dataset.shape

(250, 18)


dataset['decade']=(10 * (dataset['year'] // 10)).astype(str) + 's'


dataset['time'] = dataset['run_time'].apply(convert_time)
dataset[dataset['time'] < 1]


dataset.loc[dataset.name == 'The Boat', 'time'] = 149


dataset.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 250 entries, 0 to 249
Data columns (total 20 columns):
 #   Column             Non-Null Count  Dtype  
---  ------             --------------  -----  
 0   rank               250 non-null    int64  
 1   name               250 non-null    object 
 2   year               250 non-null    int64  
 3   rating             250 non-null    float64
 4   genre              250 non-null    object 
 5   certificate        250 non-null    object 
 6   run_time           250 non-null    object 
 7   tagline            250 non-null    object 
 8   budget             250 non-null    float32
 9   box_office         250 non-null    float32
 10  casts              250 non-null    object 
 11  directors          250 non-null    object 
 12  writers            250 non-null    object 
 13  return_rate        220 non-null    float32
 14  runtime_minutes    249 non-null    float64
 15  genre_score        250 non-null    float64
 16  certificate_score  250 non-null    float64
 17  tagline_score      250 non-null    float64
 18  decade             250 non-null    object 
 19  time               250 non-null    int64  
dtypes: float32(3), float64(5), int64(3), object(9)
memory usage: 36.3+ KB


df_unique = dataset.nunique()
ax = df_unique.sort_values().plot(kind='bar', width=0.7, sort_columns=True, ylim=(0, 1.1* df_unique.max()), color='purple')
ax.figure.set_size_inches(10,3)
ax.set_title("Contagens de valores exclusivos por coluna", weight='bold')
ax.yaxis.grid()
plt.ylabel('Contagens de valor exclusivo', weight='bold')
plt.xlabel('Cabeçalho da coluna', weight='bold')
plt.xticks(rotation=45)
bar_annot(ax, '{:d}')


ax = dataset['year'].plot(kind='hist', color='purple', width=9.5, ylim=(0,60))
ax.figure.set_size_inches(10,3)
ax.set_title("Filmes produzidos por década", weight='bold')
ax.yaxis.grid()
plt.ylabel('Contagem', weight='bold')
plt.xlabel('Década', weight='bold')
plt.xticks(rotation=45)
bar_annot(ax,'{:.0f}')
plt.show()


df_cash = dataset[['decade', 'budget', 'box_office']].groupby(['decade']).sum().reset_index()

fig, axs = plt.subplots(ncols=2, figsize=(18,5))

for i, cc in enumerate(['budget', 'box_office']):    
    ax = df_cash.plot.bar(x='decade', y=cc, color='purple', width=0.7, ax=axs[i], rot=45)
    ax.set_title(cc + " - por década", weight='bold')
    ax.yaxis.grid()
    ax.set_ylabel(cc + ' (bilhão $)', weight='bold')
    ax.set_xlabel('Década', weight='bold')
    bar_annot(ax)
fig.tight_layout()
plt.show()


df_agg = dataset[['certificate','budget','year']].groupby(by=['certificate','year']).sum().reset_index()

plt.figure(figsize=(14,5))

ax = sns.lineplot(data=df_agg, x='year',y='budget', hue='certificate', marker='o', palette=sns.color_palette('dark', n_colors=15))
ax.yaxis.grid()
ax.xaxis.grid()
ax.set_ylabel('Orçamento (bilhão $)', weight='bold')
ax.set_xlabel('Ano', weight='bold')
ax.set_title('Orçamentos de filmes por certificado', weight='bold')

plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
plt.xticks(rotation=90)
plt.tight_layout()
plt.show()


df_returns = dataset[['rank', 'return_rate', 'budget', 'box_office']]

fig, axs = plt.subplots(ncols=3, figsize=(18,5))

for i, cc in enumerate(['budget', 'box_office','return_rate']):    
    ax = df_returns.plot.scatter(x='rank', y=cc, color='purple', ax=axs[i])
    ax.set_title(cc + " - Contra Classificação", weight='bold')
    ax.yaxis.grid()
    ax.set_ylabel(cc, weight='bold')
    ax.set_xlabel('Classificação', weight='bold')
fig.tight_layout()
plt.show()


df_collab = dataset.copy() # Make a deep copy for special modifications

# Here we explode out the lists of directors, writers and casts into individual rows, duplicating all other column info
# This will be useful for analysing collaborations between people later on
for x in ['casts', 'writers', 'directors']:
    # Remove whitespace between names
    # Remove things in parentheses in names e.g. SomeName(co-director)
    df_collab[x] = df_collab[x].str.replace(' ','').str.replace(r"\(.*\)","").str.split(',').to_frame()
    # Explode to individual rows
    df_collab = df_collab.explode(x)
#df_collab.head()


fig, axs = plt.subplots(nrows=3, ncols=2, figsize=(14,15))

colors = ['purple','blue','green']

for i, cc in enumerate(['directors','casts','writers']):    
    df_agg = df_collab[[cc, 'rank', 'name', 'budget', 'box_office']].drop_duplicates().groupby([cc]).sum().reset_index()
    for j, dd in enumerate(['budget', 'box_office']):
        ax = df_agg.sort_values(dd,ascending=False).head(10).plot.bar(x=cc, y=dd, color=colors[i], ax=axs[i,j], rot=45)
        ax.set_title(cc + " " + dd, weight='bold')
        ax.yaxis.grid()
        ax.set_xlabel(cc, weight='bold')
        ax.set_ylabel(dd + " (bilhão $)", weight='bold')
        bar_annot(ax)
fig.tight_layout()
plt.show()


df_cert = dataset[['certificate', 'budget', 'box_office']].groupby(['certificate']).sum().reset_index()

fig, axs = plt.subplots(ncols=2, figsize=(18,5))

for i, cc in enumerate(['budget', 'box_office']):
    ax = df_cert.sort_values(by=cc, ascending=False).plot.bar(x='certificate', y=cc, color='purple', width=0.7, ax=axs[i], rot=45)
    ax.set_title(cc + " - Por certificado", weight='bold')
    ax.yaxis.grid()
    ax.set_ylabel(cc + ' (bilhão $)', weight='bold')
    ax.set_xlabel('Certificado', weight='bold')
    bar_annot(ax)
fig.tight_layout()
plt.show()


df_genre = dataset.copy() # Make a deep copy for special modifications

# Here we explode out the genres into individual rows, duplicating all other column info
# Remove whitespace between names and split by comma
df_genre['genre'] = df_genre['genre'].str.replace(' ','').str.split(',').to_frame()
df_genre['genre_count'] = df_genre['genre'].apply(lambda x:len(x))

# Some movies have multiple genres attached. In this case, split the financials between each genre, assuming equal weighting
df_genre['budget_split'] = df_genre['budget'] / df_genre['genre_count']
df_genre['box_office_split'] = df_genre['box_office'] / df_genre['genre_count']

# Explode to individual rows
df_genre = df_genre.explode('genre')
#df_genre.head()


df_genre_agg = df_genre[['genre', 'budget_split', 'box_office_split']].groupby(['genre']).sum().reset_index()

fig, axs = plt.subplots(ncols=2, figsize=(18,6))

for i, cc in enumerate(['budget_split', 'box_office_split']):
    ax = df_genre_agg.sort_values(by=cc, ascending=False).plot.bar(x='genre', y=cc, color='purple', width=0.7, ax=axs[i], rot=45)
    ax.set_title(cc + " - por gênero", weight='bold')
    ax.yaxis.grid()
    ax.set_ylabel(cc + ' (bilhão $)', weight='bold')
    ax.set_xlabel('Gênero', weight='bold')
    bar_annot(ax)
fig.tight_layout()
plt.show()


ax = dataset.plot.scatter(x='rank', y='rating', color='purple')
ax.figure.set_size_inches(10,5)
ax.set_title("Classificação do filme x classificação", weight='bold')
ax.yaxis.grid()
ax.xaxis.grid()
plt.ylabel('Avaliação', weight='bold')
plt.xlabel('Classificação', weight='bold')
plt.show()


df_genre_agg = df_genre[['genre','rating']].groupby(['genre']).mean().reset_index()

ax = df_genre_agg.plot.bar(x='genre', y='rating', color='purple')
ax.figure.set_size_inches(10,4)
ax.set_title("Avaliação média do filme por gênero", weight='bold')
ax.yaxis.grid()
ax.xaxis.grid()
plt.ylabel('Avaliação', weight='bold')
plt.xlabel('Gênero', weight='bold')
plt.show()


df_rate_agg = dataset[['rating', 'decade']].groupby(['decade']).mean().reset_index()

ax = df_rate_agg.plot.bar(x='decade', y='rating', color='purple')
ax.figure.set_size_inches(10,4)
ax.set_title("Avaliação média do filme por década", weight='bold')
ax.yaxis.grid()
ax.xaxis.grid()
plt.ylabel('Avaliação', weight='bold')
plt.xlabel('Década', weight='bold')
plt.show()


dataset['budget'] = dataset['budget'].astype('Int64')
dataset['box_office'] = dataset['box_office'].astype('Int64')


print("Data Size:",dataset.shape)
print("-"* 30)
print("About Dataset:")
dataset.info()
print("-"* 30)
print("Data Columns:",list(dataset.columns))
print("-"* 30)
print("Number of Examples 'N'=",dataset.shape[0])
print("Number of Dimensions 'D'=",dataset.shape[1]-1)
print("-"* 30)
print("Data Check For Any Duplicates:",dataset.duplicated().any())

Data Size: (250, 20)
------------------------------
About Dataset:
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 250 entries, 0 to 249
Data columns (total 20 columns):
 #   Column             Non-Null Count  Dtype  
---  ------             --------------  -----  
 0   rank               250 non-null    int64  
 1   name               250 non-null    object 
 2   year               250 non-null    int64  
 3   rating             250 non-null    float64
 4   genre              250 non-null    object 
 5   certificate        250 non-null    object 
 6   run_time           250 non-null    object 
 7   tagline            250 non-null    object 
 8   budget             250 non-null    Int64  
 9   box_office         250 non-null    Int64  
 10  casts              250 non-null    object 
 11  directors          250 non-null    object 
 12  writers            250 non-null    object 
 13  return_rate        220 non-null    float32
 14  runtime_minutes    249 non-null    float64
 15  genre_score        250 non-null    float64
 16  certificate_score  250 non-null    float64
 17  tagline_score      250 non-null    float64
 18  decade             250 non-null    object 
 19  time               250 non-null    int64  
dtypes: Int64(2), float32(1), float64(5), int64(3), object(9)
memory usage: 38.7+ KB
------------------------------
Data Columns: ['rank', 'name', 'year', 'rating', 'genre', 'certificate', 'run_time', 'tagline', 'budget', 'box_office', 'casts', 'directors', 'writers', 'return_rate', 'runtime_minutes', 'genre_score', 'certificate_score', 'tagline_score', 'decade', 'time']
------------------------------
Number of Examples 'N'= 250
Number of Dimensions 'D'= 19
------------------------------
Data Check For Any Duplicates: False


genre_counts = dataset['genre'].value_counts()

display(genre_counts.head(5))

real_gen = ""
for i in dataset['genre']:
    real_gen = real_gen+","+i
all_gen = real_gen.split(',')[1:]
unique_genres = list(set(all_gen))

data = []
for i in unique_genres:
    c = 0
    for j in dataset['genre']:
        if i in j:
            c+=1
    data.append([i, c])
    
gen_df = pd.DataFrame(data, columns=['Genre', 'Freq'])
gen_df.sort_values(by = 'Freq', ascending = False, inplace = True)
gen_df = gen_df.head(5)
plt.figure(figsize=(12,6),dpi = 150)
sns.barplot(data = gen_df, x = 'Genre', y = 'Freq', palette='Paired')
plt.title('Top 5 genêros mais populares')
plt.show()

---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
Cell In[1], line 1
----> 1 genre_counts = dataset['genre'].value_counts()
      3 display(genre_counts.head(5))
      5 real_gen = ""

NameError: name 'dataset' is not defined


dataset.to_csv('analise_imdb_planilha_1.csv', index=False)

	rank	name	year	rating	genre	certificate	run_time	tagline	budget	box_office	casts	directors	writers
0	1	The Shawshank Redemption	1994	9.3	Drama	R	2h 22m	Fear can hold you prisoner. Hope can set you f...	25000000	28884504	Tim Robbins,Morgan Freeman,Bob Gunton,William ...	Frank Darabont	Stephen King,Frank Darabont
1	2	The Godfather	1972	9.2	Crime,Drama	R	2h 55m	An offer you can't refuse.	6000000	250341816	Marlon Brando,Al Pacino,James Caan,Diane Keato...	Francis Ford Coppola	Mario Puzo,Francis Ford Coppola
2	3	The Dark Knight	2008	9.0	Action,Crime,Drama	PG-13	2h 32m	Why So Serious?	185000000	1006234167	Christian Bale,Heath Ledger,Aaron Eckhart,Mich...	Christopher Nolan	Jonathan Nolan,Christopher Nolan,David S. Goyer
3	4	The Godfather Part II	1974	9.0	Crime,Drama	R	3h 22m	All the power on earth can't change destiny.	13000000	47961919	Al Pacino,Robert De Niro,Robert Duvall,Diane K...	Francis Ford Coppola	Francis Ford Coppola,Mario Puzo
4	5	12 Angry Men	1957	9.0	Crime,Drama	Approved	1h 36m	Life Is In Their Hands -- Death Is On Their Mi...	350000	955	Henry Fonda,Lee J. Cobb,Martin Balsam,John Fie...	Sidney Lumet	Reginald Rose

	rank	year	rating
count	250.000000	250.000000	250.000000
mean	125.500000	1986.360000	8.307200
std	72.312977	25.125356	0.229081
min	1.000000	1921.000000	8.000000
25%	63.250000	1966.250000	8.100000
50%	125.500000	1994.000000	8.200000
75%	187.750000	2006.000000	8.400000
max	250.000000	2022.000000	9.300000

	rank	year	rating	budget	box_office	return_rate	runtime_minutes
count	250.000000	250.000000	250.000000	2.500000e+02	2.500000e+02	220.000000	249.000000
mean	125.500000	1986.360000	8.307200	4.425786e+07	2.067740e+08	inf	126.445783
std	72.312977	25.125356	0.229081	1.626479e+08	3.653295e+08	NaN	34.918079
min	1.000000	1921.000000	8.000000	0.000000e+00	0.000000e+00	0.002094	2.000000
25%	63.250000	1966.250000	8.100000	1.222000e+06	5.375080e+05	195.337830	106.000000
50%	125.500000	1994.000000	8.200000	9.750000e+06	4.603973e+07	482.502869	126.000000
75%	187.750000	2006.000000	8.400000	3.300000e+07	2.637577e+08	1108.456787	145.000000
max	250.000000	2022.000000	9.300000	2.400000e+09	2.799439e+09	inf	238.000000

	rating	name
genre
Action,Adventure	8.4	Indiana Jones and the Raiders of the Lost Ark
Action,Adventure,Comedy	8.1	The General
Action,Adventure,Drama	9.0	The Lord of the Rings: The Return of the King
Action,Adventure,Fantasy	8.7	Star Wars: Episode V - The Empire Strikes Back
Action,Adventure,Mystery	8.3	North by Northwest

Tratamento de Dados¶

O orçamento médio de um filme no conjunto de dados¶

Qual é a receita média de bilheteria de um filme no conjunto de dados?¶

Qual é o tempo de execução médio de um filme no conjunto de dados?¶

Quais diretores têm mais filmes no conjunto de dados?¶

Intervalo de valores que temos no dataset¶

Podemos encontrar o filme com a classificação mais alta em cada gênero¶

Porcentagem de filmes com base no ano¶

Bilheteria por Gênero¶

Bilheteira por Certificado¶

Agrupar anos em décadas¶

Exportar o arquivo .csv¶